Memory_initialization_radix = 16; memory_initialization_vector = 23f4 0721 11ff ABe1 0001 1 0A 0 23f4 0721 11ff ABe1 0001 1 0A 0 23f4 721 11ff ABe1 0001 1 A 0 23f4 721 11ff ABe1 0001 1 A 0; ***** ***** Example of Distributed Arithmetic FIR Filter.COE file *** *****; Example of a Distributed Arithmetic (DA) FIR Filter.COE file; with hex coefficients, 8 symmetrical taps, and 12-bit; coefficients.
![Matlab Matlab](http://xilinx.eetrend.com/files-eetrend-xilinx/blog/201706/11533-30437-fdatooldao_chu_lu_bo_qi_coe-2.png)
If you had created your BRAM memory using the Block Memory Generator you would have gotten a simulation file that already has the data as a set of (I think) generics called INITx = xxxxxxxxxx.xxxxxxxxxxxxx that have the data you had read in from a.coe file when generating the core, otherwise you would use something similar to what is described, which looks like what you are doing with Altera already as it appears your intention is to infer your RAM. $readmemh('meminitvlog.mif', mem, 0, 255); is not the synthesizable statement while.coe is. $readmemh('meminitvlog.mif', mem, 0, 255); is not the synthesizable statement while.coe is Wow, cherry pick statements and claim I'm suggesting the above? So why did you point this out like I was suggesting to use this? FYI, In some instances readmemh is synthesizable just look at both Xilinx's and Altera's synthesis guidelines. Here is an AR that describes using $readmemh to intialize an inferred block ram. Mif and coe are just two file formats that describe the same thing, the contents of a memory, whether or not they are useful for synthesis is dependent on whether they are compatible with whatever the synthesis tool requires.
$readmemh requires a file of hex values with all addresses defined and I believe no headers are allowed. The.coe file is a block ram generator tool input file that has header information besides all the binary data for the RAM based on the address locations given by the.coe file it is also NOT synthesizable, but is converted into those INITx statements I mentioned previously.